Download new zealand`s gross national product: 1859-1939

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
NEW ZEALAND’S GROSS NATIONAL PRODUCT: 1859-1939
Keith Rankin, University of Auckland, New Zealand
_______________________________________________________________
ABSTRACT
In this paper, the author continues the work of Hawke, who used Australian velocity of
money data to estimate New Zealand's GDP for 1870-1918, and whose results have been
incorporated into international studies through the work of Bairoch. He also provides an
alternative set of estimates for the inter-war years to those published by Lineham. The
important findings of the paper are: (i) that Australian data show a significant relationship
between the velocity of money and the price level; (ii) that New Zealand's income was
significantly higher in 1870 than Hawke's estimates suggest; (iii) that sustained per capita
growth has not been New Zealand's normal experience; (iv) that previous GDP estimates
for the inter-war period have failed to reflect the fluctuations of the New Zealand economy
and the extent to which it was operating below its production possibilities frontier during
the Great Depression of the 1930s.
________________
Note: These estimates were first presented to the New Zealand Association of Economists Conference,
University of Auckland, on 20 August, 1990. I would like to thank Gary Hawke, Jacques Poot, Brian Philpott,
Brian Easton, Tony Endres, Brendan Thompson, William Coleman, Geoff Bertram, Brad Patterson, Grant
Fleming and Graeme Snooks for their comments on earlier drafts of this paper.
________________
I.
INTRODUCTION
In 1962 Noel Butlin published estimates of Australia's gross domestic product (GDP),
gross national product (GNP), and gross fixed capital formation (GFCF), covering the
period from 1861 to the first official estimates of GNP in the year to June 1939. He also
published price deflators for GDP and GFCF. Unfortunately Butlin's work did not extend to
New Zealand, despite the fact that before Australian federation in 1901 New Zealand was
an Australasian colony of Great Britain. The only published time series of New Zealand's
gross product before 1918 is the GDP series of Hawke (1975). For the inter-war years, the
only complete series of gross product or national income is that of Lineham (1968). The
estimates presented here cover the 80 years of New Zealand's first century prior to the first
official GNP statistic which was published for the year to March 1939.1
Hawke, assuming comparability between the New Zealand and Australian financial
systems, used Leff’s (1972) method to construct GDP estimates from monetary data. This
paper utilises Hawke's technique as a means of interpolating between (and extrapolating
from) independently derived national income estimates which are available from
contemporary sources for a few specific years, the earliest of which is 1865. Hawke based
his estimates on the Quantity of Money Identity:
MV = PQ = Y;
1.
New Zealand was annexed as a British colony, via the Treaty of Waitangi, in 1840.
2
where M = Quantity of Money, V = Velocity of Circulation of Money, P = Price Level, Q =
Real Gross Product, and Y = Gross Product.
When a value for velocity is available, a country's income can be computed from the
size of its money stock. Since velocity data for Australia was available,2 Hawke assumed
that Australian velocity trends would be about the same as New Zealand's, given the
closeness of the two economies and the shared banking system. As a check he derived
independent velocity estimates for New Zealand for the years 1918-1933, based on the GDP
estimates of Lineham, and regressed them against Australian velocity estimates for the
same years.3 Hawke's GDP estimates were broadly in line with national income estimates
produced by contemporaries between 1886 and 1903, but incompatible with an 1865
estimate to which Hawke (1985, pp.76-77) nevertheless gives a considerable degree of
credence. Comparisons with Australian incomes suggest that Hawke's data understate New
Zealand's GDP before the 1880s and after the mid-1900s. Hawke's series is also suspect
because New Zealand's fluctuations were not well synchronised with Australia's (Dowie,
1963).
While Hawke was careful to state that his 1975 estimates of New Zealand's GDP were
no better than "plausible" (1985, p.79) or a "stop-gap" (1975, p.306), his series has
nevertheless been utilised by national and international studies (Fairburn, 1989,
pp.98,106-9; Bairoch, 1981, p.10; De Long, 1988, p.1152).4 Bairoch’s 1860 and De Long’s
1870 statistics are serious underestimates of New Zealand's GDP. De Long's 1870 New
Zealand per capita datum is only half of the value given for Australia. In the light of the
2.
3.
4.
Velocities were calculated by dividing Butlin's GDP estimates by monetary aggregates (bank deposits)
provided by S.J. Butlin, Hall and White (1971, pp.140-156).
Hawke (1975, p.304) obtained the formula: log VNZ = log 1.54 + 0.19 VA (R = 0.52).
Bairoch only gives a reference to an unpublished paper, but it is apparent from an earlier paper,
(1977, p.185) that his 1860 New Zealand datum is estimated from Hawke's series. De Long cites Bairoch
as his source.
3
estimates in this paper, New Zealand is an excellent supporting example for Baumol's
(1986) "convergence" hypothesis which De Long's study challenges.
II. VELOCITY REGRESSIONS
In trying to discover why Hawke's GDP estimates for the 1870s were so low, it
became apparent that the Australian velocity estimates might be correlated to the general
price level. With New Zealand prices falling more rapidly than Australian prices from the
1860s to the 1880s, it seemed likely that Hawke's velocity approximations for the 1870s
were too low.5 An estimation function for velocity could be improved, evidently, by
including a price variable. Further examination of the Australian data showed that velocity
was not particularly high in the late 1880s, the period of "Marvellous Melbourne" (Davison,
1978) in which Victoria's economy was characterised by a speculative investment boom. A
reason for the low velocity statistic during the boom becomes apparent from a consideration
of the quantity identity.
Strictly speaking, MV = PT, where "T" is the volume of transactions rather than
output. Thus:
MV = PT
<=>
MV = PQ.(T/Q)
<=>
V' = PQ/M
where V' = V.(Q/T) = Y/M.
Here, V' is regarded as the ratio of gross national product to trading bank deposits.6
This velocity ("VEL") can be expected to be comparatively low at times in which there is
much trade in existing assets. Such periods, when the volume of transactions is high
relative to national income, are indicative of a high speculative demand for money.7 The
general price level also tends to rise at such times.
It is therefore proposed that the quantity of money per capita is negatively related to
its velocity. The essence of Hawke's technique is that the monetary data is scaled up by a
velocity parameter to produce GDP estimates. With velocity inversely related to the stock
of money, the derived gross product data fluctuate less than the quantity of money data.
This paper’s estimates are based on estimating velocity ("VEL") from two regressors:
trading bank deposits per capita ("MPC") and the price level ("PRI"). Hawke's and
Lineham's GDP series are graphed in Figure 1, alongside this paper's GNP estimates.
Australia's GNP is presented in Table 1 and Figure 2.8
... Table 1 here ...
The regression equations are derived from the Australian data presented in Table 1.9
They exclude the years 1914-18, and include a dummy variable ("IW") for the inter-war
5.
6.
7.
8.
9.
Hawke (1985, pp.78-79) noted that he had, in effect, assumed that velocity decreases were less marked in
New Zealand than in many other countries during the 19th century.
I have chosen to estimate gross national product at market prices, rather than gross domestic product, in
order to link up with the official GNP estimates for 1938/39.
Kindleberger (1984, p.365) describes such a situation: "while the overall money supply is unchanged,
within the total there is a shift from what Keynes called the 'transactions circulation' to the 'financial
circulation'" which is the same as "a decline in income velocity".
Butlin's (1962, pp.6-7) estimates, scaled to link with official estimates for 1938-40 (Butlin 1962, p.468),
and converted to calendar years.
All time series relating to financial years have been converted to calendar years. I have used Butlin's
price series (1962, pp.33-34) and monetary data from Butlin, Hall and White (1971, pp.142-57). The
4
years 1919-39. There were 74 observations. Three equations have been estimated, using
respectively the current year's ("MPC"), the previous year's ("MPC-1"), and both years' per
capita monetary aggregates. The equations found are, with t-statistics in brackets:
(a) ln VEL = -7.94 + 1.57 * ln PRI - 0.602 * ln MPC - 0.429 * IW
[12.1] [17.5]
[20.8]
[6.97]
2
R = 0.944
DW = 0.99
F = 394
(b) ln VEL = -7.27 + 1.47 * ln PRI - 0.608 * ln MPC-1 - 0.371 * IW
[11.0] [16.3]
[21.0]
[5.93]
2
R = 0.945
DW = 1.46
F = 402
(c) ln VEL = -7.52 + 1.51 * ln PRI - 0.27 * ln MPC - 0.34 * ln MPC-1 - 0.39 * IW
[11.2] [16.4]
[1.64]
[2.01]
[6.20]
2
R = 0.947
DW = 1.23
F = 309
The statistical relationship between the price level and velocity is very strong, as is the
negative relationship between the quantity and velocity of money. Although the lagged
money stock is a slightly better regressor than current-year stock, a more robust function for
New Zealand's velocity can be constructed by including both lagged and current "MPC"
series, giving them equal weight. It is also clear from the dummy variable that structural
changes in the relationships between the regression variables occurred during the war.
These differences between the inter-war years and the pre- World War I years raise further
doubts about Hawke's adoption of a function to estimate velocity based on inter-war
monetary data.
Australian population data - averages of the December 31 estimates - is from Maddock and McLean
(1987, pp.353-54) and the Official Year Book of the Commonwealth of Australia (1910, pp.118-19).
5
New Zealand velocities are likely to reflect some structural factors that do not apply to
Australia, because of possible differences in the demand for banking services due to factors
such as variation in income distribution and economic geography. These differences are not
likely to have been the same in 1939 as they were in 1859. Thus, more accurate New
Zealand velocity estimates can be found by using separate regression equations for separate
periods, and by scaling the resulting estimates to fit independently derived New Zealand
benchmarks. The 80 year time-frame was divided into three sub-periods, and separate
regressions on the Australian data were taken for each.
The periods chosen for the three regressions were 1861-1900, 1900-13, and 1919-39.
The fitted equations are as follows:
(d) 1861-1900
ln VEL = -6.10 + 1.33 * ln PRI - 0.190 * ln MPC - 0.491 * ln MPC-1
[5.75] [9.23]
[0.858]
[2.15]
2
R = 0.964
DW = 1.48
F = 320
(e) 1900-1913
ln VEL = -7.46 + 1.32 * ln PRI + 0.097 * ln MPC - 0.238 * ln MPC-1
[2.21] [2.08]
[0.058]
[0.478]
R2 = 0.670
DW = 1.28
F = 6.76
(f) 1919-1939
ln VEL = -8.35 + 1.47 * ln PRI - 0.594 * ln MPC + 0.167 * ln MPC-1
[8.52] [14.5]
[2.20]
[0.592]
2
R = 0.935
DW = 0.74
F = 81.5
The relative importance of the current and lagged monetary aggregates varies for each
function. But when they were used as substitute rather than complementing regressors - as
in equations (a) and (b) - the coefficients were similar whichever monetary series was used.
Serial correlation is present. The effects are shown in Figure 2. In the 1920s, the lack of
randomness of the deviations is marked but their amplitudes are generally small. There are
a few short periods where the predicted Australian GNP deviates by about 10 percent from
the Table 1 series. This suggests that Australia experienced some events for which dummy
variables would raise the regression R2; periods such as the early 1860s, 1879-81,
mid-1890s, early 1920s. Similar periods for New Zealand must also have occurred. A likely
candidate would be the late-1880s - analogous to the mid-1890s in Australia - when GNP
may have been significantly lower than has been estimated.
III. BENCHMARK ESTIMATES
Independent estimates for New Zealand's gross national product have been derived for
the years 1865, 1898/99-1902/03, 1925/26, 1932-33, and 1938/39.
6
______________________________________________________
TABLE 2: Benchmark Estimates of New Zealand's GNP.
Year
GNP (£m)
1865
15.8
1898/99
36.8
1900/01
43.3
1902/03
46.2
1925/26
175.7
1932
117.0
(1932/33
113.7)
1933
123.7
1938/39
231.1
______________________________________________________
The first benchmark is a contemporary estimate of national income by Charles Knight
(1866), a senior public servant. Dowie (1966), in assessing Knight's estimate of £15.8
million, has concluded that any errors which may have lead to overstatement appear to be
balanced by factors which would lead to understatement.10 Dowie concluded that Knight's
effort was one of great intellectual merit, overcoming a lack of data and an absence of
methodological precedent. His national income total has therefore been taken as a valid
estimate of New Zealand's GNP for 1865. Knight's work indicated that New Zealanders'
incomes were significantly higher than Australians' which in turn were well above
prevailing incomes in any other country. This conclusion has to be seen, however, in the
context of higher prices in New Zealand, the low population base, and by the fact that
average incomes had been temporarily boosted by the gold rushes of 1861/62 and 1865/66.
The data shown in Figure 2 suggests that there was income parity between Australia and
New Zealand in 1860 and 1870.
Dowie also cited four estimates of New Zealand's turn of the century GNP by the New
South Welsh economist and statistician, T.A. Coghlan: 1898/99, 1900/01, 1901/02 and
1902/03.11 Coghlan’s achievements have been assessed by Arndt (1949) and Butlin (1962,
pp.36-40). Dowie raised Coghlan's Australian estimates by 12.5 percent to make them
comparable with Butlin's GDP estimates.12 The benchmarks used for the years 1898/99,
1900/01, and 1902/03 are the adjusted Coghlan estimates.
Since the introduction of the Reserve Bank in 1934 created a discontinuity in New
Zealand's monetary data, it is inappropriate to apply the monetary technique beyond 1933.
Instead, I use semi-official estimates of GNP from 1931/32 which were reconstructed from
10.
Dowie discusses Knight’s procedures in some detail, finding that some “double-counting” in the primary
and banking sectors gives an error of about 10 percent of the total estimate. However an allowance for
house rents reduces the error to 3 percent. Knight himself believed his estimate was an understatement,
because he omitted some activities for which he could find no data.
11. Coghlan’s estimates were for Australia and New Zealand, before and after Australian federation in 1901,
disaggregated into colonies/states.
12. Butlin's GNP estimates are lower than his estimates of GDP at market prices. However, the Table 1 GNP
estimate for 1902 is equal to Butlin’s GDP estimates for 1901/02 and 1902/03, so the 12.5% scale factor
still applies.
7
aggregate private income data published in the 1940s.13 The 1930s’ benchmarks are taken
from this series. Note that the trough of New Zealand’s Great Depression was the year to
March 1933, somewhat later than most countries, with an estimated GNP of only £113.7
million. The 1938/39 benchmark is the first official estimate (NZOYB, 1957, p.716).
Fisher (1930) estimated aggregate private income for 1925/26 from census and tax
data. His estimate was £137.1 million. Assuming that the Fisher estimate of aggregate
private income is comparable with the earliest official estimates for the years 1931/32 to
1935/36 (NZOYB, 1938, p.737), to convert to GNP his estimate has been scaled up by the
same amount that these first official estimates have been scaled. Since estimates of GNP are
produced in this paper, and because the 1932 benchmark coincides with the peak of the
Great Depression, Fisher's estimate was further adjusted in line with the ratio of GNP:GDP
in Butlin's Australian data. In the middle of the Depression in Australia, GNP was
unusually low compared to GDP. The scaled Fisher estimate (£175.7m) serves as a useful
benchmark, which can be checked against Lineham's GDP estimate of £168.8m for
1925/26. Lineham's series is likely to be comparatively accurate for that year because it was
a census year, and because it was a year of low unemployment.14 Lineham's 1938/39 GDP
estimate is 95 percent of the official GNP statistic. Similarly, his 1925/26 estimate is 96
percent of the Fisher benchmark.
IV. THE NEW ZEALAND DATA SERIES
Because prices constitute a key regressor used to determine velocity, it is necessary to
have a series of prices which can be used with some confidence. A price deflator is also
necessary to estimate the real GNP per capita series needed to evaluate economic growth
and to compare New Zealand and Australian income levels. The deflator used, presented in
Table 3 along with bank deposits and population, is a linked series of four segments. The
principal segments are the wholesale price index for locally produced products for
1913-40,15 and a conflation of four price indexes (McIlraith, 1911; Easton & Wilson, 1984)
for 1861-1910: imports, exports, farm, non-farm. The link from 1910 to 1913 is made up by
combining the official retail and wholesale price indexes (NZOYB). The 1859-60 price
estimates are derived from the three series provided by Easton and Wilson: imports, all
exports, and farm exports. The final deflator series has been based on 1910/11, to conform
with the Australian series given in Table 1. Purchasing power parity between the two
countries has been assumed for the base years. The monetary data in Table 3 is the series of
trading bank deposits given by Bloomfield (1984, pp.386-87). The population data is
exclusive of indigenous Maoris.16 The import/export data derives from Bloomfield
(pp.267-269)17.
... Table 3 here ...
13.
14.
15.
16.
17.
NZOYB (1957, p.717; 1940, pp.782-83; 1942, p.647). I adjusted the series from March years to calendar
years so as to make them compatible with the regression estimates. Note also the GNP/GDP series in
NZOYB 1990 (p.679) which goes back to 1935.
Lineham is likely to have overstated GDP for 1935/36, another census year, because in years of high
unemployment there was always a degree of underemployment, making wage rates an unreliable guide to
wage incomes.
NZOYB 1941 (p.710). Note other price indexes for the inter-war period in Rankin (1990a, p.40), and
that the chosen inter-war series is near to an average of these.
Year-end averages from the NZOYB (1901 p.501, 1919 p.849, 1924 p.726) series, and from the New
Zealand Statistics of Population and Buildings for each year 1922/23-1939/40.
Bloomfield’s series have been deflated by this paper’s GNP deflator. Thus, the series depicted in
Figure 3 represent trade values, not volumes.
8
The functions used to estimate New Zealand velocities are constructed from the
Australian regressions (equations (d) to (f) ). Equal weighting is given to the current and
lagged monetary aggregates. The New Zealand velocity and GNP estimates are presented in
Table 4. The New Zealand equations are:18
(g) 1859-1897:
VEL = 0.98 * exp [-6.10 + 1.33 * ln PRI + (-0.190 + -0.491) * avg (ln MPC, ln MPC-1) ]
(h) 1895-1913:
VEL = 1.18 * exp [-7.46 + 1.32 * ln PRI + (0.097 + -0.238) * avg (ln MPC, ln MPC-1) ]
(i) 1922-1933:
VEL = 1.18 * exp [-8.35 + 1.47 * ln PRI + (-0.594 + 0.167) * avg (ln MPC, ln MPC-1) ]
The estimates for 1895-97 are made from weighted averages of equations (g) and (h).
Those years, chosen partly for the pragmatic reason that equation (h) gave a good fit to the
Coghlan benchmarks, constituted a key turning point in the economic history of New
Zealand. The "Long Depression" had just come to an end. Coinciding with the Liberal
Government's land and labour reforms of 1893-94, the opening up of the North Island dairy
industry, made viable by the introduction of refrigerated shipping, gave people with access
to capital the confidence to raise investment spending. One result was an increase in the
velocity of money without price stability being compromised. It is therefore appropriate that
velocity estimates from 1895 should be higher than they would have been under equation
(g).
... Table 4 here ...
Velocity for the World War I years has been estimated arbitrarily, with the aim of
producing GNP estimates which are consistent with other information about economic
activity. The war period is something of an enigma in New Zealand's economic history.
New Zealand and Australia both seem to have had a recession that was disguised by
inflation and the fact that both countries' labour surpluses were in Europe. In contrast to
World War II, the number of factories and factory workers in New Zealand fell
(AJHR H-11, 1919, p.4) and the volume of farm production was static. Condliffe
(1924/25, p.231) claimed that "during the war and ... immediately following, production fell
off considerably in New Zealand". He describes the period around 1920 as a "post-war
boom", and it is clear that a strong multiplier effect, which was not in operation in 1918,
was boosting New Zealand's GNP in 1920. There was no war-time labour shortage despite
the big fall in labour supply.19 The Labour Department Report (AJHR H-11, 1918, p.1) for
1917/18 states that "despite anticipations to the contrary, there were more men available
during harvest time than were required". An increase in female employment between the
1916 and 1921 censuses appears to be related to falling real disposable household incomes
in the face of inflation; that is, an increase in female labour supply rather than in any
demand for substitute female workers. Indeed, the 1918/19 Report (AJHR H-11, 1919, p.5)
18.
For example, the coefficients of (g) are exactly the same as those for equation (d). However, the
coefficients for "MPC" and "MPC-1" are added together to give a single coefficient, which is applied to
the average of those two variables. The number "0.98" is the scale factor used to fit the equation to the
1865 New Zealand benchmark.
19. New Zealand’s manpower losses to World War I were very large (Neale, p.76).
9
indicates that the increase in teenage factory employment at the end of the war coincided
with falls in overtime.
Equation (i), when applied to 1919-21, gives an unrealistically high GNP, especially
for 1920. The New Zealand economy did not fully settle into its new pattern until 1922.
Post-war money growth was much sharper in New Zealand than Australia. It was linked to
land speculation as well as restocking and fixed capital formation. For those years the
predicted velocity values have been adjusted to give more plausible estimates. That is
somewhat arbitrary, but gives a boom/bust phase for 1918-22 that is sharper than 1905-09
and less pronounced than 1876-79. This conforms with the general impressions of these
periods in the historical literature.
V. DISCUSSION OF THE ESTIMATES
The new estimates of real GNP per capita, compared to contemporary estimates of
national income cited by Dowie (1966, p.127) and Lineham (1968, p.25), are shown in
Figure 1. The 1886 estimate of £30 million by Otago University's Professor Mainwaring
Brown fits the GNP estimate of this paper almost exactly. Dowie (1966, pp.130-31)
regarded Brown's work as an important contribution, although not of the same significance
as Knight's.
Mulhall produced estimates for 1888 and 1895 for inter-country comparisons for the
London Dictionary of Statistics (1892, 1909). He used two formulae which are of some
interest (Dowie, 1966, n15), but his own estimates derived from those techniques cannot be
taken with any degree of precision because the data used were taken from a variety of
different years. For example, for his 1895 estimate, data were taken from various years in
the early 1890s. Mulhall's 1895 estimate for Australia of £179 (Butlin 1962, p.37) exceeds
the Australian GNP (given in Table 2) for 1891-95 by 5 percent, while his New Zealand
estimate of £34.2 for 1895 exceeds the Table 4 New Zealand average for those years by a
similar 4 percent. Therefore, this paper’s GNP estimates for the early 1890s are fully
consistent with his 1895 estimate of Australia's gross product. Mulhall's 1888 aggregate of
£22.5m appears to be well short of New Zealand's true GNP for that year, although the GNP
estimate given here is probably too high. 1888 was the year of the New Zealand "Exodus";
the trough of the Long Depression and a period of mass emigration to Australia.
In the 1894 and 1897 Official Year Books, the New Zealand Registrar-General
presented national income estimates using both the Brown and Mulhall methods.20
Averaging his two estimates gives £27.5 million for 1893, and £28.6 million for 1896. With
both numbers being about 18 percent short of the estimates given in this paper, the
Registrar-General's estimates consistently understate them.21
Lineham follows a direct approach to the task of gross product estimation for the
inter-war period, but his estimates have serious empirical weaknesses (Rankin 1991,
pp.15-18). He presents contemporary national income estimates by Stephens (1936) and
Clark.22 Stephens' estimates for 1925/26-1930/31, which are too low for 1925, have been
included in Figure 1. Financial, labour market and trade data indicate that the New Zealand
20.
Dowie (1966,p.127); NZOYB (1894, p.139; 1897, pp.283-284)
As the Registrar-General's 1896 estimate is based on data from the census in April 1896, to make the
comparison I adjusted my estimates to the year ended September, by taking weighted averages of that
year and the previous year. His estimates are 18% short of my estimates for the year to September 1893
and the year to September 1896. This slight adjustment is significant because GNP recovered by over 13
percent in 1896.
22. Lineham (p.25) has demonstrated a high level of correlation (R 2=0.97) between Stephens’ and Clark’s
(2nd. edn., 1951) estimates. Clark (1957, p.173) sources his estimates to an unpublished memorandum by
Stephens and himself.
21.
10
economy was buoyant in 1925 but in sharp recession by 1927. Stephens' estimates for
1928-30 are close to those of this paper, although they show a faster slide into depression in
1930. His data are for years ending the following March, so his income estimate for 1930
will have been considerably affected by the much greater extent of depression in early 1931
than in early 1930.
To be plausible, a GNP series constructed by indirect means should be consistent with
other macroeconomic data. Changes in the real value of New Zealand's import payments
(shown in Figure 3) indicate the major fluctuations, with the growth in export receipts
acting as a leading indicator of economic growth. Australia-New Zealand migration data
show that the New Zealand GNP estimates are fully comparable with the Australian series.
An inflow of Australian migrants to New Zealand would be expected to follow periods of
higher GNP or economic growth in New Zealand. The data in Figure 4 shows that New
Zealand experienced a net inflow of Australian migrants in years in which New Zealand
incomes were higher (1860s, 1893, 1900s, 1929-31, 1937-39), and net emigration to
Australia in years of comparatively low GNP (1880s, 1926-28, 1933-35).23 For most years,
per capita incomes in the two countries were similar. A reduced level of contact between
Australia and New Zealand after 1908 can be explained by the general similarity of living
standards and of comparative advantage.
23.
The Gandar adjustments, which have been converted to calendar years, assume that the entire
discrepancy for the years in question related to Trans-Tasman migration. Figure 4b only covers people
claiming to be permanent migrants. This is because returning New Zealanders ceased to be classified by
place of origin.
11
It is important to note that the per capita GNP data are exclusive of the indigenous
Polynesian (Maori) population. A census of Maori population in 1858 (Bloomfield,
pp.37,42) showed that Maoris comprised 43.5 percent of the total population. That
proportion had fallen to 13 percent when the next Maori census was taken in 1874.24
Knight’s 1865 estimate of national income was based solely on the European population.
New Zealand’s total GNP would have been much higher than stated for 1860 if the full
contribution of the Maori economy had been accounted for. As an example of the level of
24.
The Maori population fell from 56,000 in 1858 to less than 40,000 in 1896 before it began to rise.
12
integration in 1860, the Maori warrior and prophet Te Kooti was then a coastal trader on the
Auckland-Gisborne route, using two Maori-owned ships to undercut the former settler
monopoly (Oliver 1990, p.463). As late as 1880, after much of the best Maori-owned land
had been confiscated, the lands around the Taranaki town of Parihaka were farmed
productively by best-known techniques (Riseborough 1989, p.130). Nevertheless, from the
Land Wars in the 1860s until the 1940s, there was little integration between the indigenous
and settler economies.
The peaks of per capita income in the 1860s coincided with the gold discoveries in
Otago (1861), Westland (1865) and Coromandel (1868). The New Zealand economy was
also growing strongly because of the high demand in Britain for wool, New Zealand’s main
staple. Expansion in the 1870s was boosted by a program of Government sponsored public
works and immigration (the "Vogel boom") financed by borrowing on London. In the
1880s, the government and many producers were burdened by indebtedness in the face of
falling land and commodity prices. In the absence of an export staple with powerful
linkages into the domestic economy, New Zealand tentatively moved towards a
manufacturing-based economy based in part on the availability of cheap female and teenage
labour (Rankin 1990b, pp.11-13; Sutch, 1957, p.25 n41), and also on the buoyancy of the
Australian market. In the 1890s, the economy was boosted by a new set of staple primary
industries - the processing of meat and dairy products - made possible by advances in
refrigeration technology and the growth of the British market for imported food.
A feature of the New Zealand aggregates from the 1890s which was not apparent in
Australia is the existence of a 3-5 year trade cycle. From 1907 to 1929, the fluctuations
were very sharp oscillations around a nil-growth trend. These short period cycles were the
main source of the uncertainty and "instability" that pervaded New Zealand's economic
consciousness (Simkin, 1951; Hawke, 1985, p.100). They were the price New Zealand paid
for its narrowly structured economic development, based on pastoral exports to the British
market. Those fluctuations explain the popular desire for an insulationist approach to
economic policy-making in the years following World War II. The sharpness of some of the
single-year fluctuations in GNP reflect the nature of investment activity in a small
commodity-producing economy. For the whole period, the domestic money supply was
essentially determined by the level of reserves held by the New Zealand banking system in
London. Thus, rapid growth could result from an acceleration of domestic investment made
possible by an increase in London funds. A sharp fall in the British price of one or more of
New Zealand’s primary export commodities, or a change of sentiment towards Australia or
New Zealand by British investors, would inevitably lead to a sharp contraction of the New
Zealand money supply. The 1879 downturn was linked to the collapse of the City of
Glasgow Bank in 1878 (Hawke, 1985, p.80) and the 1926 recession was mainly a result of
the General Strike in Great Britain.
Economic growth rates were very high for most of the 1930s, averaging as much as
8 percent from 1933 to 1938. Lineham (1968, p.18) noted that "the massive public works
program would probably have set the economy off on an exponential rather than a linear
growth path". This multiplier effect is also the result of the earnings of large numbers of
additional workers, mainly women and teenagers, who sought work during the Depression
and who contributed to a rapid increase in household incomes during the recovery years
(Rankin, 1990a, p.123).
VI. CONCLUSION
In this paper I have presented a set of GNP estimates for New Zealand that can be
linked to official estimates beginning in 1938/39. They draw heavily on the work of
13
contemporary economic statisticians. Although no formal model linking monetary
aggregates to GNP has been presented, it has been possible to produce a gross product
series that stands up to historical scrutiny. An important by-product of this exercise has
been the demonstration of a significant statistical relationship between prices and monetary
velocity in the Australian data.
It should be possible to produce improved historical estimates of New Zealand's
national income by following procedures similar to those used by Butlin's team in Australia.
Indeed, the statistical and archival record has yet to be mined in New Zealand to the extent
it has in many other developed countries. The GNP estimates therefore remain a provisional
data source against which future estimates can be checked. Their virtue is that they give a
better indication of economic growth in New Zealand’s first century than any alternative
source. They present a macroeconomic picture containing long periods of nil trend growth;
a picture which gives credence to the considerable anxiety felt by New Zealanders about
their country's economic destiny. That pattern continues today. Per capita incomes are now
little higher than they were in 1975, and many economists doubt that a growth trend will
emerge before 1995.
______________
14
APPENDIX
A. THE ECONOMIC SIGNIFICANCE OF THE REGRESSION VARIABLES
While the concept of monetary velocity provides the means to link monetary and
income/product variables, in this paper I use variables contained within the quantity identity
to estimate velocity. Thus it is possible to remove velocity entirely from the equation,
leaving a society's gross product as a function of the money stock, the price level and the
population.
From equation (a) we have
(A1) ln V = a + b.lnP - c.ln(M/N)
b > 1; 0 < c < 1;
(A2) V = A.Pb.(M/N)-c
where V = velocity
ln A = a
P = prices
M = money stock
N = population
From the Quantity Identity we have
(A3) PQ = M.V
(A4) Q/N = M.V.P-1.N-1
where PQ = gross product
Q = real gross product
Taking the two together:
(A5) Q/N = A.Pb-1.(M/N)1-c
(A6) Q/N = A.Pb-c.(M/P/N)1-c
(A7) Q = A.Pb-1.M1-c.Nc
where M/P/N = real money balances per capita
Looked at in this way, equation (A7) shows that real gross product is positively related
to population, money balances and prices. An increase of 1 percent in any of those three
variables will, in itself, be associated with an increase of less than 1 percent in gross
product. Equation (A6) suggests that rising real money balances are linked with economic
growth only when prices are rising.
It is not possible to claim a simple causal link from prices, money or population to real
incomes because of the interdependence of the variables. For example, a sharp rise in prices
may induce a significant fall in real money balances. Rather, the above is a historical
generalisation with respect to the Australian economy, and ipso facto, the New Zealand
economy. Economic growth in New Zealand resulted from market situations which brought
about both rising prices and a growth in the money supply.
B.
THE RELATIONSHIP OF THE QUANTITY AND VELOCITY OF
MONEY TO PRICES
As a by-product of the regressions conducted for this exercise, I performed another set
of regressions on the full Australian data set (1861-1939, excl. 1914-18), this time making
the price level the independent variable. The data yielded the following equations (with
t-values in brackets):
(B1) ln PRI = 7.22 - 0.0995 * ln MPC-1 + 0.604 * IW
[63.9]
[2.7]
[16.2]
R2 = 0.888
15
(B2) ln PRI = 6.65 + 0.252 * ln VEL-1 + 0.548 * IW
[215.5] [9.0]
[33.9]
R2 = 0.942
(B3) ln PRI = 5.40 + 0.314 * ln MPC-1 + 0.529 * ln VEL-1 + 0.32 * IW
[45.5] [10.7]
[17.0]
[13.6]
R2 = 0.979
As the results of equation (B1) suggest, if there is a simple causal relationship from
money (narrowly defined) to prices, then it is negative. This result is contrary to that
predicted by monetarist doctrine. Instead, there is a strong positive relationship from
velocity to prices in equation (B2). This is consistent with the discussion at the beginning
with this paper about the relationships between monetary velocity and prices, and between
the velocity of and the quantity of money. A significant direct relationship from money to
prices is revealed, however, when velocity is accounted for as a separate regressor, as in
equation (B3).
--------------------
REFERENCES
Appendices to the Journal of the New Zealand House of Representatives (AJHR).
Arndt, H.W., “A Pioneer of National Income Estimates”, Economic Journal 59, pp.616-25,
1949.
Bairoch, P., "Estimations du Revenu National dans les Sociétés Occidentales
Pré-Industrielles et au Dix-Neuvième Siècle: Propositions d'Approches Indirectes",
Revue Economique, pp.177-208, 1977.
Bairoch, P. and Lévy-Leboyer, M. (eds.), Disparities in Economic Development Since the
Industrial Revolution., St. Martins Press, New York, 1981.
Baumol, W.J., "Productivity, Growth, Convergence and Welfare", American Economic
Review 76, pp.1072-85, 1986.
Bloomfield, G.T., A Handbook of New Zealand Statistics, Hall & Co., Boston, 1984.
Butlin, N.G., Australian Domestic Product, Investment and Foreign Borrowing,
1861-1938/39; Cambridge University Press, Cambridge, 1962.
Butlin, S.J., Hall, A.R., and White, R.W., Australian Banking and Monetary Statistics
1817-1945, Reserve Bank of Australia (Occasional Paper no.4A), 1971.
Clark, Colin, The Conditions of Economic Progress (3rd. edn.), Macmillan, London, 1957.
Condliffe, J.B., "Changing Price Levels", The Accountants Journal of New Zealand, vol.3,
pp.139-46, 197-202, 229-232, 1924/25.
Davison, G.R., The Rise and Fall of Marvellous Melbourne, Melbourne University Press,
Melbourne, 1978.
De Long, Bradford, "Productivity Growth, Convergence and Welfare: Comment",
American Economic Review 78, pp.1138-1154, 1988.
Dowie, J. A., "Inverse Relations of the Australian and New Zealand Economies,
1871-1900", Australian Economic Papers 2, pp.151-179, December 1963.
16
Dowie, J. A., "A Century-Old Estimate of the National Income of New Zealand", Business
Archives and History (now Australian Economic History Review) 6, pp.117-131,
August 1966.
Easton, B., and Wilson, N., "An Investigation of the Database of New Zealand's
Terms of Trade", the First Draft of a Report Prepared by the New Zealand Institute
of Economic Research for the Department of Trade and Industry, Working Paper
no. 84/10, Appendix 2, 1984.
Fairburn, Miles, The Ideal Society and its Enemies: the Foundations of Modern New
Zealand Society, 1850-1900; Auckland University Press, Auckland, 1989.
Fisher, A.G.B., "Distribution of Income in New Zealand", The Economic Record 11,
pp.221-234, 1930.
Gandar, J.M., “New Zealand net migration in the latter part of the nineteenth century”,
Australian Economic History Review 19, pp.151-168, 1979.
Hawke, G.R., "Income Estimation from Monetary Data: Further Explorations", Review of
Income and Wealth 21, pp.301-07, 1975.
Hawke, G.R., The Making of New Zealand, Cambridge University Press, Cambridge, 1985
Kindleberger, C.P., Financial History of Western Europe , Allen and Unwin, London, 1984
Knight, Charles et al., "Final Report of the Civil Service Commissioners",
AJHR 1866 D-7B.
Leff, N.H., "A Technique for Estimating Income Trends from Currency Data and an
Application to Nineteenth Century Brazil", Review of Income and Wealth 18,
pp.355-68, 1972.
Lineham, B.T., "New Zealand's Gross Domestic Product, 1918/38", New Zealand
Economic Papers 2, pp.15-26, 1968.
McIlraith, J.W., The Course of Prices in New Zealand, Govt. Print, Wellington, 1911.
Maddock, Rodney & McLean, I.W., The Australian Economy in the Long Run; Cambridge
University Press, Cambridge, 1987.
Neale, E.P., "The Post-War Sex- and Age-Constitution of the New Zealand Population and
some of its Economic Consequences", Economic Record 8, pp.75-82, 1929.
New Zealand Official Year Book (NZOYB), various issues.
New Zealand Statistics for Population and Buildings, official publication, 1922/23 1939/40.
New Zealand Statistics of External Migration, official publication, 1920s.
Official Year Book of the Commonwealth of Australia (OYBCA), 1910.
Oliver, W.H. (ed.), Dictionary of New Zealand Biography 1769-1869 (vol.1), Allen and
Unwin, for the New Zealand Department of Internal Affairs, 1990.
17
Rankin, Keith, Labour Supply in New Zealand and Australia, 1919-1939; unpublished
thesis, Victoria University of Wellington, 1990a.
Rankin, Keith, "New Zealand’s Labour Supply in a Long Term Perspective", Working
Paper 5/90, Industrial Relations Centre, Victoria University of Wellington, 1990b.
Rankin, Keith, "Gross National Product Estimates for New Zealand; 1859-1939", Graduate
School of Business and Government Management, Working Paper Series 1/91,
Victoria University of Wellington, 1991.
Riseborough, Hazel, Days of Darkness; Taranaki 1878-1884, Allen and Unwin,
Wellington, 1989.
Simkin, C.G.F., The Instability of a Dependent Economy: Economic Fluctuations in New
Zealand, 1840-1914, Oxford University Press, Oxford, 1951.
Statistics of New Zealand, official publication, (various years).
Sutch, W.B., The Long Depression, Industries and Commerce Department, Wellington,
1957.
Stephens, F.B., "National Income of New Zealand", Economic Record 27, pp.231-56, 1936
==================